You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 6eac1cb
Browse filesBrowse the repository at this point in the historyBrowse files
Copy file name to clipboardExpand all lines: docs/plans/2026-10-08-impact-gate-rollout.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -36,7 +36,7 @@ The upstream baseline serializes `_meta` (`tool`, `n`, `head`, `base_ref`) plus
36
36
37
37
## Actions verification still required
38
38
39
-
Local calibration and tests cannot verify GitHub cache scoping, cache saves, later PR restores, or the rendered job summary. Any run that rebuilds the pair saves it, pull requests included; a PR's cache is scoped to its own ref, so it cannot replace main's, and it spares later pushes to that PR a full rebuild. Main pushes run in per-SHA concurrency groups so a quick second merge cannot cancel the first before it saves. After the workflow is available in Actions, record links to a cold run, a main run saving the pair, a PR run restoring it, and a PR run reusing its own saved pair. Verify both report sections, policy/version cache invalidation, malformed-pair rebuilding, and explicit seed-only disclosure. Do not mark these checks complete based on local cache simulation.
39
+
Local calibration and tests cannot verify GitHub cache scoping, cache saves, later PR restores, or the rendered job summary. Any run that rebuilds the pair saves it, pull requests included; a PR's cache is scoped to its own ref, so it cannot replace main's, and it spares later pushes to that PR a full rebuild. Each save's key ends with its run ID and attempt, because cache keys are immutable: a pair rebuilt after the one restored for the same commit failed validation saves beside it instead of being refused, and the next run restores the newer one. Main pushes run in per-SHA concurrency groups so a quick second merge cannot cancel the first before it saves. After the workflow is available in Actions, record links to a cold run, a main run saving the pair, a PR run restoring it, and a PR run reusing its own saved pair. Verify both report sections, policy/version cache invalidation, malformed-pair rebuilding, and explicit seed-only disclosure. Do not mark these checks complete based on local cache simulation.
0 commit comments